Coronary CT in Australia has high positive predictive value unaffected by site volume: An analysis of 510 positive CTCA scans with invasive angiographic correlation

Abstract

BackgroundIt was hypothesized that the accuracy of coronary CT angiography would be affected by case volume of the referring sites. MethodsThe positive predictive value (PPV) of CTCA performed at a tertiary hospital specialising in cardiothoracic medicine and services with lower case-volumes were calculated. The tertiary hospital used as the high case-volume reference centre was The Prince Charles Hospital, which performed >1500 CTCA scans per annum over the study period. The low case-volume services used in the study were suburban radiology services, each with <500 cases per year. The PPV of positive CTCA at the reference site was compared to the pooled PPV of all other sites as a combined cohort, using invasive angiography as the reference standard. 512 scans were included, n = 199 subjects in the reference centre cohort, and n = 311 subjects in the pooled community radiology practice cohort. ResultsThe positive predictive value (PPV) of the high case-volume group (n = 199) was 0.7538. The PPV of the pooled low case-volume services (n = 589) was 0.7331, p = 0.604, with no statistically significant difference in positive predictive values. ConclusionsThere was no significant difference in PPV between the two groups. This suggests that high-volume and lower-volume sites both have high PPV in Australia, above the published pooled PPV of four large prospective diagnostic accuracy studies (Miller et al., 2008; Budoff et al., 2008; Meijboom et al., 2008; Achenbach, 2007).